MCP Catalogs
Homemcp_flutter screenshot

mcp_flutter

by Arenukvern·294·Score 50

MCP server for Flutter AI agents to inspect, interact with, and control running Flutter applications.

developer-toolsai-llmother
32
Forks
14
Open issues
this month
Last commit
2d ago
Indexed

Overview

This is a Dart-based MCP server and Flutter toolkit that enables AI assistants to interact with running Flutter applications. It provides 27 MCP tools under the 'fmt_*' prefix for app inspection, interaction, debugging, and lifecycle management. A key feature is the ability to dynamically register custom tools and resources at runtime, creating a closed feedback loop between AI agents and Flutter apps.

Try asking AI

After installing, here are 6 things you can ask your AI assistant:

you:AI-driven Flutter app testing and debugging
you:Automated UI interaction and form filling in Flutter apps
you:Visual regression testing using AI agents
you:How do I connect the MCP server to my Flutter app?
you:Can I create custom MCP tools for my Flutter app?
you:Which AI assistants are supported?

When to choose this

Choose this when working with Flutter apps and need AI agents to interact with UI elements, take screenshots, or debug running applications through a closed feedback loop.

When NOT to choose this

Avoid if you need to interact with native iOS/Android code directly, or if your workflow requires more general-purpose app testing tools not specific to Flutter.

Tools this server exposes

12 tools extracted from the README
  • fmt_list_client_tools_and_resources

    List all client tools and resources available in the Flutter app

  • fmt_hot_reload

    Trigger a hot reload of the Flutter app

  • fmt_hot_restart

    Trigger a hot restart of the Flutter app

  • fmt_semantic_snapshot

    Take a semantic snapshot of the Flutter app's UI

  • fmt_tap_widget

    Tap on a specific widget in the Flutter app

  • fmt_type_into_text_field

    Type text into a specific text field in the Flutter app

  • fmt_scroll

    Scroll the Flutter app view in a specific direction

  • fmt_screenshot

    Take a screenshot of the current Flutter app view

  • fmt_get_recent_logs

    Retrieve recent logs from the Flutter app

  • fmt_evaluate_dart_expression

    Evaluate a Dart expression in the Flutter app context

  • fmt_navigate_to_route

    Navigate to a specific route in the Flutter app

  • fmt_wait_for_condition

    Wait for a specific condition to be met in the Flutter app

Comparable tools

flutter-toolsappiumdetoxflutter-driver

Installation

Installation

  1. Install the binary:
curl -fsSL https://raw.githubusercontent.com/Arenukvern/mcp_flutter/main/install.sh | bash
  1. Add the toolkit to your Flutter app:
cd my-flutter-app
flutter-mcp-toolkit codegen-init   # adds flutter_mcp_toolkit + emits main.dart snippet
  1. Install skills for your AI agent:
flutter-mcp-toolkit init claude-code   # or: cursor | codex | cline | all
  1. Run your app:
flutter run --debug

Claude Desktop Configuration

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "flutter-mcp-toolkit": {
      "command": "flutter-mcp-toolkit",
      "args": ["server"]
    }
  }
}

FAQ

How do I connect the MCP server to my Flutter app?
Ensure your Flutter app is running in debug mode. The toolkit provides a WebSocket connection that the MCP server can connect to. You can find the connection URI in your Flutter app's debug output.
Can I create custom MCP tools for my Flutter app?
Yes, the toolkit supports dynamic tool registration at runtime. You can add custom tools using `MCPToolkitBinding.instance.addEntries()` in your Flutter app.
Which AI assistants are supported?
The toolkit supports Claude Code, Cursor, Codex, Zed, Intent, Cline, and other compatible AI assistants that can connect to MCP servers.

Compare mcp_flutter with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.